Recent breakthroughs in nucleic acid sequencing technology have made possible the sequencing of entire genomes from a variety of organisms, including humans. The potential benefits of a complete genome sequence are many, ranging from applications in medicine to a greater understanding of evolutionary processes. These benefits cannot be fully realized, however, without an understanding of how and where these newly sequenced genes function.
Traditionally, functional understanding started with recognizing an activity, isolating a protein associated with that activity, then isolating the gene, or genes, encoding that protein. The isolated protein was also used to generate antibody reagents. Specific antibodies and fragments of the isolated gene were both employed to study tissue expression and function.
Several methods have been used to study protein expression patterns including in situ hybridization studies of tissue sections and northern blots. These methods are both time consuming and require relatively large amounts of material to perform successfully.
Antibodies that bind to specific antigens have been produced by a variety of methods including immunization of animals, fusion of mammalian spleen cells to immortalized cells to produce hybridomas, random peptide generation using phage or bacterial display and constrained peptide libraries. Regardless of how the desired antibody is generated, the methods currently available to identify one with a particular binding specificity are generally laborious and incapable of the simultaneous testing of large numbers of unknowns.
One method involves binding the antigen to a porous membrane, such as nitrocellulose, contacting the membrane with a source of test antibodies, then determining whether or not any of the test antibodies has bound to the antigen. This method only allows the testing of one source of test antibodies per piece of porous membrane, making the method both inconvenient and wasteful of materials.
Antibody/antigen reactions can also be evaluated in plastic plates, such as 96-well microtiter plates, using methods similar to those described above. This method is likewise limited in the number of samples that can be tested in any one assay, thus requiring many assays to fully evaluate a large number of antibody unknowns. Chang (U.S. Pat. No. 4,591,570, issued May 27, 1986) describes an array of a limited number of characterized antibodies to known antigens on a glass surface that can be used to bind to specific antigens on the surface of whole cells.
Recently new technologies have arisen that allow the creation of microarrays containing thousands or millions of different elements. Such array technology has been applied mainly to forming arrays of individual nucleic acids (see, for example, Marshall and Hodgson, Nature Biotech. 16:27-31, 1998; Ramsay, Nature Biotech. 16:40-44, 1998), in particular short oligonucleotides synthesized in situ.
Methods are needed to simply and rapidly screen very large numbers of uncharacterized antibodies for those specific for a given antigen as well as for the characterization of tissues and cells by nucleic acid and/or protein analysis. The invention described herein addresses that need.